Publication Type

Journal Article

Version

publishedVersion

Publication Date

11-2021

Abstract

While literature has reached a consensus on the awareness effect of online word-of-mouth (eWOM), this paper studies its persuasive effect, specifically, the dimension-specific sentiment effects on product sales. We allow the sentiment information in eWOM along different product dimensions to have different persuasive effects on consumers’ purchase decisions. This occurs because of consumers’ sentiment preference, which is defined as the relative importance consumers place on various dimension-specific sentiments. We use an aspect-level sentiment analysis to derive the dimension-specific sentiments and PVAR (panel vector auto-regression) models to estimate their effects on product sales using a movie panel dataset. The findings show that three dimension-specific sentiments (star, genre and plot) are positively related to movie sales. Regarding consumers’ sentiment preferences, we find that the positive relationship with movie sales is stronger for plot sentiment relative to star sentiment for low-budget movies, whereas the positive relationship with movie sales is stronger for star sentiment relative to plot and genre sentiments for high-budget movies.

Keywords

Online Word of Mouth, Dynamic Topic Model, Sentiment Analysis, Product Sales

Discipline

Databases and Information Systems | E-Commerce | Sales and Merchandising

Research Areas

Information Systems and Management

Publication

Journal of the Association for Information Systems

Volume

22

Issue

2

First Page

459

Last Page

489

ISSN

1536-9323

Identifier

10.17705/1jais.00668

Publisher

Association for Information Systems

Embargo Period

7-12-2021

Copyright Owner and License

Authors

Additional URL

https://doi.org/10.17705/1jais.00668

Download

Share

COinS